Genetic Programming and Redundancy   [GP]

by

Blickle, T. and Thiele, L.

Literature search on Evolutionary ComputationBBase ©1999-2013, Rasmus K. Ursem
     Home · Search · Adv. search · Authors · Login · Add entries   Webmaster
Note to authors: Please submit your bibliography and contact information - online papers are more frequently cited.

Info: Genetic Algorithms within the Framework of Evolutionary Computation (Workshop at KI-94, Saarbrücken) (Conference proceedings), 1994, p. 33-38
Keywords:genetic algorithms, genetic programming
Notes:
From GP list Wed, 22 Mar 95 we did some work on the convergence problem and the redundancy in the trees in GP. It turned out that "bloating" is a property of GP that arises from the fact that more redundant trees have a higher probability to survive crossover. As a result, the redundant part of the trees grow bigger and bigger because the increased proportion of redundant "cut-sites" in the tree again lead to a higher probability to survive crossover. Gives a formula for tournament size related to proportion of crossover in a generational GP. Ie recommending T=10 for pc=0.9. This does not apply to steady state GA.
URL(s):(G)zipped postscript

Review item:

Mark as doublet (will be reviewed)

Print entry



BibTex:
@InProceedings{BT94,
  author =       "Tobias Blickle and Lothar Thiele",
  title =        "Genetic Programming and Redundancy",
  booktitle =    "Genetic Algorithms within the Framework of
                 Evolutionary Computation (Workshop at KI-94,
                 Saarbr{\"u}cken)",
  editor =       "J. Hopf",
  publisher =    "Max-Planck-Institut f{\"u}r Informatik
                 (MPI-I-94-241)",
  address =      "

                 Im Stadtwald, Building 44, D-66123 Saarbr{\"u}cken,
                 Germany

                 ",
  pages =        "33--38",
  year =         "1994",
  keywords =     "genetic algorithms, genetic programming",
  URL =          "http://www.tik.ee.ethz.ch/~blickle/GPandRedundancy.ps.gz",
  size =         "6 pages",
  notes =        "From GP list Wed, 22 Mar 95 we did some work on the
                 convergence problem and the redundancy in the trees in
                 GP. It turned out that {"}bloating{"} is a property of
                 GP that arises from the fact that more redundant trees
                 have a higher probability to survive crossover. As a
                 result, the redundant part of the trees grow bigger and
                 bigger because the increased proportion of redundant
                 {"}cut-sites{"} in the tree again lead to a higher
                 probability to survive crossover.

                 Gives a formula for tournament size related to
                 proportion of crossover in a generational GP. Ie
                 recommending T=10 for pc=0.9. This does not apply to
                 steady state GA.

                 ",
}